JPH01224843A - Tool for expert system construction with operating mode - Google Patents

Abstract

PURPOSE:To raise a processing speed on operating by providing two modes which are an operating mode and a debugging mode to a tool for expert system construction, making them switchable, and passing a checking processing or the like for debugging at the operating mode. CONSTITUTION:When an instruction to switch to a debugging mode side is inputted from an external part, a mode control part 1-7 switches a mode flag 1-7-1 to the debugging mode side, and the same action as usual is executed on the development of an expert system. On the other hand, when the development of the expert system is completed and the instruction to switch to an operating mode is inputted from the external part, the mode control part 1-7 switches the mode flag 1-7-1 to the operating mode side. Thus, a knowledge base processing part and an inferential engine pass a processing on debugging and execute only a processing on operating, and a processing speed is improved for a portion in which respective checking processing or the like for debugging are passed.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a tool for constructing an expert system, and in particular, an expert system with an operational mode that improves processing speed by streamlining main processing in the operational environment of the expert system. Concerning construction tools.

[Conventional technology]

An expert system is an expert system that stores the knowledge of experts in a knowledge base and uses the knowledge stored in this knowledge base to cause an inference engine to perform inferences to solve problems in a specific field. This is a system that allows ordinary people without knowledge to make judgments equivalent to those of experts.

Various methods have been proposed or put into practical use for developing such expert systems, but an expert system construction tool is used as a method that allows the development of highly practical expert systems in a relatively short period of time. There is a method.

Expert system construction tools include the knowledge base, inference engine, knowledge base processing section, user ink face section, and
Provides a knowledge debugger, a knowledge editor, etc., stores various knowledge necessary for the expert system to be developed using the knowledge editor in the knowledge base, debugs it using the knowledge debugger, etc., and further debugs it in advance. By performing operations such as selecting an appropriate inference method for the expert system from among the various inference methods provided, the expert system construction tool itself can be used as an expert system for solving a specific problem. This is a system that allows you to

FIG. 4 shows the procedure for developing an expert system using conventional expert system construction tools. First, experts or knowledge engineers (hereinafter referred to as KE)
), the necessary knowledge is input into the knowledge base using the knowledge editor (step 4-1). Next, start the system and debug it (Step 4-
2).

At this time, since the knowledge base processing unit and inference engine that constitute the expert system construction tool are equipped with check processing and the like to support debugging, debugging is proceeded based on the results of the check processing and the like.

If an error is found as a result of the execution, the knowledge is corrected using the knowledge editor (step 4-3). When debugging is completed (step 4-4), the development process is completed, and the expert system is now in operation by the user (step 4-5).

Stages of expert system operation (steps 4-5)
In this case, the inference engine accesses various types of knowledge in the knowledge base via the knowledge base processing unit in response to various questions input by the user via the user interface unit.
Actions are performed to draw a conclusion by performing various inferences and the like. Conventional expert system construction tools provide the same internal environment for the development stage for debugging and the operational stage for using the completed expert system.

[Problem that the invention seeks to solve]

As described above, conventional expert system construction tools provide the same internal environment for the development stage for debugging and the operation stage for using the completed expert system.

That is, various check processes that are necessary only at the development stage are actually performed even at the operational stage.

This was one of the causes of slow processing speed during operation.

Therefore, an object of the present invention is to provide an expert system construction tool with two modes, an operation mode and a debug mode, and to enable switching between them.In the operation mode, by passing check processing for debugging, etc. The objective is to provide tools for building expert systems with increased speed.

[Means for solving problems]

In order to achieve the above object, the present invention includes a knowledge base that stores various types of knowledge, a knowledge base processing unit that manages access to the knowledge base, and a knowledge base that accesses the knowledge base via the knowledge base processing unit. In an expert system construction tool that includes an inference engine that executes inference for problem solving while applying various acquired knowledge, and a user interface section between the inference engine and external input/output devices, The knowledge base processing unit and the inference engine include a mode control unit that switches a mode flag to an operation mode side or a debug mode side in response to an input mode setting instruction; In addition to the operation time processing and the operation time processing for executing the inference, there is a configuration that performs the debug time processing necessary for debugging, and when the mode flag is set to the operation mode side, the debug time It is configured to pass the process and execute only the operation process, and when the mode flag is set to the debug mode side, to execute the operation process while performing the debug process.

[Effect]

When an instruction to switch to the debug mode side is input from the outside, the mode control unit switches the mode flag to the debug mode side. As a result, the knowledge base processing unit performs debugging processing to perform various checks, etc., and also performs operational processing to manage access to the knowledge base.
The inference engine also executes inference while performing debugging processing that performs various checks. Therefore, when developing an expert system, the same operations as before are performed. - On the other hand, when the development of the expert system is completed and an instruction to switch to the operational mode side is input from the outside, the mode control unit switches the mode flag to the operational mode side. As a result, the knowledge base processing unit and the inference engine bypass the debugging process and execute only the operational process, and the processing speed is improved by the amount that various check processes for debugging are passed.

〔Example〕

Next, embodiments of the present invention will be described with reference to the drawings.

FIG. 1 is a block diagram of an embodiment of the present invention. As shown in the figure, the expert system construction tool T of this embodiment
L is knowledge base l-1, knowledge base processing unit 1-2. Inference engine 1-3. It consists of a user ink face section 1-4, a knowledge editor 1-5, a knowledge debugger 1-6, and a mode control section 1-7 having a mode flag 1-7-1, and input/output devices 1-9 and 1-. 10.

Knowledge editor! -5 is connected to the input/output device 1-10 and the knowledge base processing unit 1-2, and manually inputs various kinds of knowledge necessary for the expert system to be realized by this expert system construction tool TL into the knowledge base t-t. Provides a man-machine interface when With this knowledge editor l-5, you can easily add knowledge to the knowledge base 1-1.
can be entered.

The knowledge debugger 1-6 is connected to the input/output device 1-10 and the knowledge base processing unit 1-2, and has various debugging functions effective for debugging the knowledge stored in the knowledge base 1-1. provide.

The knowledge base processing unit 1-2 includes a knowledge base 1-1, an inference engine l-3, a knowledge editor l-5, and a knowledge debugger 1-.
6 and the mode control unit 1-7, the knowledge base 1
In this embodiment, the knowledge base processing unit 1-2 manages the overall access to the 1-1 knowledge base processing that is executed when the mode flag l-7-1 in the mode control unit 1-7 indicates the debug mode. Debug mode section 1-2-1
and a knowledge base processing operation mode section 1-2-2 included therein and executed when the mode flag 1-7-1 indicates the operation mode. The knowledge base processing debug mode unit 1-2-1 provides general processing for accessing the knowledge base 1-1 during expert system development.

Therefore, various check processes necessary for debugging are also included. The knowledge base processing operation mode section 1-2-2 provides general processing for accessing the knowledge base 1-1 during operation, and the knowledge base processing debug mode section 1-2
-1 minus the check processing necessary for debugging. That is, the knowledge base processing operation mode section 1-2
-2 can be said to be the core part of the knowledge base processing unit 1-2.

Inference engine l-3 is a knowledge base processing unit! -2, is connected to the user interface section 1-4 and the mode control section 1-7, and can solve problems while using various knowledge obtained by accessing the knowledge base l-1 via the knowledge base processing section 1-2. In this embodiment, the inference engine 1-3 executes inference for a solution by setting the mode flag l- in the mode control unit 1-7.
An inference engine debug mode section 1-3-1 that is executed when 7-1 indicates the debug mode, and an inference engine operation mode included therein that is executed when the mode flag 1-7-1 indicates the operation mode. It consists of part l-3-2. The inference engine debug mode section 1-3-1 provides a function of separating various inference methods and actually proceeding with inference according to the methods. The inference engine operation mode section 1-3-2 controls inference during operation, and is the section of the inference engine debug mode section 1-3-1 except for check processing necessary for debugging. . That is, the inference engine operation mode units 1-3 and -2 can be said to be the core of the inference engine 1-3.

The knowledge base 1-1 is a means for storing expert knowledge, and is connected to the knowledge base processing section 1-2. Generally, a knowledge base stores factual knowledge, rule knowledge, and the like.

The mode control section 1-7 includes a user interface section 1-4.
．． It is connected to the knowledge base processing unit 1-2 and the inference engine 1-3, and has an internal mode flag 1-7-1. This mode control section 1-7 has a user interface section 1-7.
When a mode flag setting command is input via 4, the mode flag 1-7-1 is set according to the content of the command.
Switch to debug mode or operational mode.

The user ink face section 1-4 is connected to the mode control section 1-7.
．． It is connected to the inference engine 1-3 and the input/output device 1-9, and provides a man-machine interface when the expert system is executed. Further, during debugging at the development stage, various commands are input from this user interface section 1-4 to proceed with debugging.

FIG. 2 shows an example of the configuration of the knowledge base processing section 1-2. Generally, the knowledge base processing unit 1-2 includes a plurality of processes A, B,
. . . to control access fees to the knowledge base 1-1. That is, processing A, processing B,...
is an essential process for the knowledge base processing unit 1-2. Also,
For debugging, it is necessary to perform some kind of check process a, b, etc. for every process A, process B, . . . or for each necessary process among them. The conventional knowledge base processing unit 1-2 performs these check processes a and b.
, . . . were configured to be executed also during operation of the expert system, but in this embodiment, each process A.

Immediately after B, .
For example, only when F=1), check processing etc. a, b, .
..., and when the operation mode is indicated (for example, when F=0), those check processes, etc. a,
b, . . . are configured to be bused.

The inference engine 1-3 is also configured similarly to the knowledge base processing section 1-2. That is, when the inference engine 1-3 performs processing necessary for debugging such as check processing after the processing essential for inference, it determines the mode flag 1-7-1, which indicates the debug mode side. It is configured to perform check processing etc. only when the

FIG. 3 shows the procedure for developing an expert system using the expert system construction tool with operation mode of the present invention. First, an expert or KE inputs necessary knowledge into the knowledge base 1-1 using the knowledge editor 1-5 (step 3-1), and then inputs the necessary knowledge into the knowledge base 1-1 through the user interface section 1-4. By inputting a command to instruct switching to the operational mode side, the mode control unit 1
The mode flag 1-7-1 in -7 is switched to the operational mode side (step 3-2). Then, the system is actually started and debugged (step 3-3). If the mode flag 1-7-1 is set to the debug mode side in this way, the user interface section 1
In response to the command input from -4, inference engine 1-
3, the inference engine debug mode section 1-3-1 is executed, the knowledge base processing section 1-2 executes the knowledge base processing debug mode section 1-2-1, and each debug mode section 1-3-1.1 Since the check processing provided in -2-1 is executed, debugging can proceed in the same manner as before. If there is an error as a result of execution, the knowledge is corrected using the knowledge debugger 1-6 (step 3-5), and when debugging is completed (
Step 3-4), the development process is completed.

Before putting the developed expert system into operation,
Input an instruction to switch to the operation mode side via the user interface section 1-4, and switch the mode flag l-7-1 of the mode control sections 1 to 7 to the operation mode side (
Step 3-6), and then the process proceeds to the stage of operation of the expert system by the user (step 3-7).

At the stage of operation of the expert system (step 3-7), the inference engine l-3
accesses various kinds of knowledge in the knowledge base 1-1 via the knowledge base processing section 1-2, and performs various inferences and the like to draw a conclusion. In this embodiment, when the mode flag 1-7-1 indicates the operation mode side, the inference engine operation mode section 1-3-2 is executed in the inference engine 1-3, and the knowledge base processing section 1-3 is executed.
In 2, the knowledge base processing operation mode section 1-2-2 is executed, and the check processing etc. in the inference engine debug mode section 1-3-1 and the knowledge base processing debug mode section 1-2-1 are bused. . In this way, the processing speed during operation of the expert system is increased.

In the embodiment shown in FIG. 1, switching of the mode flag 1-7-1 was performed by a command input from the user interface section 1-4, but the knowledge editor 1-5. The switching may be performed using a command input from the knowledge debugger 1-6, or a mode flag may be provided in the knowledge base processing unit 1-2 and the inference engine 1-3.

【Effect of the invention〕

As described above, according to the present invention, it is possible to improve the processing speed during operation of an expert system developed using an expert system construction tool.

[Brief explanation of the drawing]

Figure 1 is a block diagram of an embodiment of the present invention, Figure 2 is a diagram showing an example of the configuration of the knowledge base processing unit 1-2, and Figure 3 is an expert system construction using the expert system construction tool of the present invention. FIG. 4 is a diagram showing the procedure for constructing an expert system using a conventional expert system construction tool. In the figure, TL... Expert system construction tool 1-1.
...Knowledge base 1-2...Knowledge base processing section 1-2-1...Knowledge base processing debug mode section 1-
2-2...Knowledge base processing operation mode section 1-3...
Inference engine 1-3-1...Inference engine debug mode section 1-3
-2...Inference engine operation mode section 1-4...User interface section 1-5...Knowledge editor 1-6...Knowledge debugger 1-7...Mode control section 1-7- 1...Mode flag 1-8.1-9...Input/output device patent applicant 1 person other than NEC Corporation

Claims (1)

[Claims] A knowledge base that stores various kinds of knowledge, a knowledge base processing section that manages access to the knowledge base, and a knowledge base processing section that manages access to the knowledge base through the knowledge base processing section. In an expert system construction tool that includes an inference engine that executes inference for problem solving while applying the application, and a user interface section between the inference engine and external input/output devices, mode settings input from the outside are used. a mode control unit that switches a mode flag to an operating mode side or a debug mode side in response to an instruction, and the knowledge base processing unit and the inference engine are configured to perform operation processing for managing access to the knowledge base and the In addition to operational processing to perform inference,
It has a configuration that performs debugging processing necessary for debugging, and when the mode flag is set to the operation mode side, the debugging processing is passed and only the operation processing is executed, and the mode flag is set to the operating mode side. A tool for constructing an expert system with an operational mode, characterized in that when the tool is set to a debug mode side, the operational process is executed while performing the debugging process.